Objective: To analyze the expression of B-cell development-related genes in acute B lymphoblastic leukemia (B-ALL), and to explore the relationship between B-cell development-related genes and the prognosis of B-ALL patients.
Methods: The GEO and TARGET databases were integrated to analyze the differential expression of B-cell development-related genes between the healthy persons and B-ALL patients and their differential expression in the B-ALL relapse and non-relapse groups. Cox single factor regression and Lasso regression were used to constructe a B-ALL specific prognosis model of B-cell development-related genes. The prognostic value of this model was analyzed by Cox multiple factor regression. The risk scores of different subtypes of B-ALL was analyzed. In the real world, the correlation between the prognostic model of B-cell development-related genes and clinical outcomes was verified through the transcriptome sequencing results of B-ALL patients. In addition, the correlation between this prognostic model and other B-ALL prognostic models was also analyzed. At last, Metascape was used to evaluate the pathway and function enrichment status related to the prognosis model.
Results: There were 1 097 genes specifically expressed in B-ALL and related to B cell development, 27 of which were up-regulated in the B-ALL relapse group, and 37 genes were down-regulated in the B-ALL relapse group. 14 genes were further selected to be included in the B-cell development-related prognosis model (CDC25B,CKAP4,DSTN,IGF2R,NDUFA4,ODC1,PAX5,SH3BP4,SLC27A5,APAF1,ARRB2,HHEX,IL13RA1,UVRAG) based on Cox single factor regression and Lasso regression. Risk scoring of patients with B-ALL based on the 14 genes prognosis model, the prognosis of 134 patients in the low-risk scoring group (score>0.11) was better than those in the patients with high-risk scores (score≤0.11). Multivariate analysis showed that the risk score of B-cell development-related genes was an independent prognostic factor. And the proportion of hyperdiploid positive children in the low-risk scoring group was significantly higher than that in the high-risk scoring group, while the proportion of TCF3/PBX1 positive children in the high-risk scoring group was significantly higher than that in the low-risk scoring group. At the same time, the real-world data showed that the prognosis of patients with B-ALL in the high-risk scoring group was worse than those of the patients with low-risk scores in Xi'an Children's Hospital. And the risk score of B-cell development-related genes in patients with B-ALL death was higher than that in patients with B-ALL non-death. In addition, there is a positive correlation between the risk score calculated by the metabolic-related gene prognostic scoring system and the risk score calculated by the B-cell developmental-related gene prognostic model. At last, differential gene enrichment analysis suggested that the prognosis risk was related to the process of embryonic development and differentiation to various systems, especially to the B cell receptor signaling pathway.
Conclusion: The specific expression of B-cell development-related genes in B-ALL is related to the prognosis of B-ALL. The prognosis model composed of 14 genes is expected to be a new prognostic marker for children with B-ALL.
题目: B细胞发育相关基因在儿童急性B淋巴细胞白血病中的表达及预后意义.
目的: 分析B细胞发育相关基因在儿童急性B淋巴细胞白血病（B-ALL）中的表达，探索B细胞发育相关基因与B-ALL患儿预后的相关性。.
方法: 基于GEO和TARGET数据库，分析健康对照者与B-ALL患儿差异表达的B细胞发育相关基因及其在B-ALL复发组和未复发组的差异表达。采用Cox单因素回归及Lasso回归方法筛选候选基因并构建B-ALL特异性的B细胞发育相关基因的预后模型。通过Cox多因素回归评估所构建的预后模型的应用价值，并分析B-ALL不同亚型的风险评分情况。在真实世界中,通过B-ALL患儿转录组测序结果验证B细胞发育相关基因预后模型与临床结局之间的相关性。此外，还分析了该预后模型与其他B-ALL预后模型的相关性。最后，采用Metascape评估与该预后模型相关基因的信号通路及功能富集状态，以探究其潜在机制。.
结果: 在B-ALL中特异性表达且与B细胞发育相关的基因共1 097个，其中27个基因在B-ALL复发组中表达上调，37个基因在B-ALL复发组中表达下调。采用Cox单因素回归及Lasso回归方法，筛选出14个基因纳入B细胞发育相关基因的预后模型（CDC25B、CKAP4、DSTN、IGF2R、NDUFA4、ODC1、PAX5、SH3BP4、SLC27A5、APAF1、ARRB2、HHEX、IL13RA1、UVRAG）。基于14个基因的预后模型对TARGET数据库中134例B-ALL患儿进行风险评分，高风险评分组（评分>0.11）的患儿预后差于低风险评分组（评分≤0.11）的患儿。Cox多因素分析显示，该B细胞发育相关基因的风险评分可作为B-ALL患儿独立预后因素，并且低风险评分组中高二倍体阳性患儿的比例显著高于高风险评分组，而高风险评分组中TCF3/PBX1阳性患儿的比例显著高于低风险评分组。同时，真实世界数据显示，高风险评分组的B-ALL患儿预后差于低风险评分组的患儿，且B-ALL死亡组患儿B细胞发育相关基因的风险评分高于B-ALL未死亡组。此外，根据代谢相关基因预后积分系统计算的风险评分与该B细胞发育相关基因预后模型计算的风险评分具有正相关性。最后，差异基因的功能富集分析提示，B-ALL患儿的预后风险与胚胎向各系统发育分化过程，尤其与B细胞受体信号通路相关。.
结论: 在B-ALL中特异性表达的B细胞发育相关基因与B-ALL患儿的预后相关，其中14个基因构成的预后模型有望成为儿童B-ALL新的预后判断标志。.
Keywords: B-cell developmentrelated genes; acute B lymphoblastic leukemia; prognosis; children.